Mint Quick Edit | Deepfake menace: Why tagging truth may work better than labelling fakes

On Wednesday, it proposed draft rules that require AI content to be labelled by its creators and social media platforms; the latter would also need to scrutinize such content for takedowns.

The menace of fakes is real and intervention might help, but policing every byte of content that goes online may simply not be feasible.

A more pragmatic approach may be to enlist market incentives for the task. We could devise a system of provenance certification for authentic content to be tested and labelled as such.

Those who want to convey the truth have an incentive to stand out for it, while those aiming to deceive us would prefer to operate below the fake-spotting radar.

This would lead to a self-sorting market, with discerning consumers of content looking out for authenticity tags. Whatever is left uncertified would lose relative credibility. So long as these tags can’t be gamed, this may offer a better solution to our problem of whether to believe what we see online or not. The idea is worth a shot.